AI Report Calls for Greater Access to Public Sector Data
A 125-page report, “Machine learning: the power and promise of computers that learn by example,” published by the Royal Society, the UK’s national academy of science, argues for greater access to public sector data for machine learning research. The report is penned by a working group composed of researchers from leading AI companies, including Google DeepMind, Amazon, and Uber.
The report highlights the potential benefits of open data for AI advancement, including increased productivity and economic growth. However, it also acknowledges the concerns surrounding the use of public data, including privacy and the potential for commercial advantage by those with early access.
The report’s authors argue that public data should be made “open by default” and structured in a way that supports “wider use of research data.” They recommend new open standards for data and suggest revisiting the concept of “consent” in the context of data governance to facilitate wider data access.
The report specifically mentions the value of NHS data as a key asset for machine learning. However, it also highlights the controversy surrounding DeepMind’s data access partnerships with the NHS, which were agreed upon without patient consent.
While the report acknowledges the potential for commercial advantage from access to public data, it refrains from explicitly addressing the potential for monopolies. The report suggests that “appropriately controlled access mechanisms” could be developed to balance public good with privacy concerns.
Overall, the report underscores the importance of open data for AI development while acknowledging the need for responsible data governance.
